Applied Scientist I, Ads Trust Science
Core
Build multimodal ML models to moderate millions of ad requests daily, ensuring ads are safe across languages and cultures for Amazon DSP customers.
Role type
Applied Scientist I (Multimodal Content Understanding)
Builds
Production ML models for ad moderation
Domain
Advertising technology, Trust & Safety, Multimodal AI
Deliverable
production ML models
Required skills
Deep learning architecture design, Multimodal model training (vision + language), NLP, Computer vision, Java, C++, Python, SQL, RDBMS (via careerplan.io/jobs/10566528-applied-scientist-i-ads-trust-science-at-amazon)
Preferred skills
Publications at top-tier conferences, Applied theoretical models in production
Technologies
Vision models, Language models, Deep learning frameworks
Responsibilities
Design and train multimodal models to flag or clear ads, Build systems generalizing across languages and cultural contexts, Deploy models from notebook to production pipelines, Partner with engineers to scale systems to Amazon scale
Seniority
Individual Contributor, Mid-level